In a prototypical ferromagnet (Ga,Mn)As based on a III–V semiconductor, substitution of divalent Mn atoms into trivalent Ga sites leads to severely limited chemical solubility and metastable specimens available only as thin films. The doping of hole carriers via (Ga,Mn) substitution also prohibits electron doping. To overcome these difficulties, Masek et al. theoretically proposed systems based on a I–II–V semiconductor LiZnAs, where isovalent (Zn,Mn) substitution is decoupled from carrier doping with excess/deficient Li concentrations. Here we show successful synthesis of Li1+y(Zn1−xMnx)As in bulk materials. Ferromagnetism with a critical temperature of up to 50 K is observed in nominally Li-excess (y=0.05–0.2) compounds with Mn concentrations of x=0.02–0.15, which have p-type metallic carriers. This is presumably due to excess Li in substitutional Zn sites. Semiconducting LiZnAs, ferromagnetic Li(Zn,Mn)As, antiferromagnetic LiMnAs, and superconducting LiFeAs systems share square lattice As layers, which may enable development of novel junction devices in the future.
Figures at a glance
Figure 1: Crystal structures of Li(Zn,Mn)As and X-ray and magnetization results. 
(a–c): crystal structures of (a) cubic LiZnAs (direct-gap semiconductor) and Li(Zn,Mn)As (ferromagnetic metal, as demonstrated in the present work), (b) tetragonal LiMnAs (antiferromagnet with TN>393K), and (c) tetragonal LiFeAs (non-magnetic superconductor with Tc~20K). All these structures include square-lattice As layers with the matching of the lattice constant within ~10%. Primitive cells for the tetragonal crystals of (b) and (c) are shown by the black lines. (d,e): X-ray scattering results of (d) intensity profile of Li1.1(Zn,Mn)As with various Mn concentrations x, and (e) lattice constant of Li1.1(Zn1−xMnx) As for various x (blue symbols, bottom horizontal axis), and Li1+yZnAs (red symbols, top horizontal axis) for various Li deficiency/excess y. (f,g): Magnetization M(H,T) results of Li1.1(Zn1−xMnx)As with x=0.0–0.15 showing (f) the T dependence of M in H=2 kOe (no difference in FC and ZFC procedures) and (g) M at T=2 K in various values of external field H. The grey symbols show a hysteresis loop in x=0.03 system plotted for small field regions (top horizontal axis), which demonstrate a very small coercive field of 30–100 Oe.
Figure 2: Results of transport measurements. 
Transport results of sintered polycrystalline specimens of Li1+y(Zn1−xMnx)As: (a) resistivity of Li1+yZnAs, showing metallic behaviour of Li deficient (y<0) and Li excess (y>0) compounds. (b) resistivity of Li1.1(Zn1−xMnx)As, showing the effect of increasing charge scattering with increasing Mn concentration x. (c) resistivity of Li1.1(Zn0.9Mn0.1)As in various external field H, which exhibits negative magnetoresistance below Tc~50 K. (d) Hall resistivity of Li1.1(Zn0.95Mn0.05)As at T=2 K, which exhibits p-type carriers with concentrations of n~1020 cm−3 together with the anomalous Hall effect due to spontaneous magnetization at H=0.
Figure 3: Results of muon spin relaxation measurements. 
Results of μSR measurements in sintered polycrystalline specimens of Li1.1(Zn0.95Mn0.05)As: (a) time spectra in zero field that exhibit onset of extra relaxation below T~30 K. The solid lines represent fits to the relaxation function for dilute spin systems in zero field for the static case (often used for dilute-alloy spin glasses13), which exhibits a fast relaxation, plus a non-relaxing paramagnetic component (Methods); (b) the relaxation rate a of the signal that exhibits fast relaxation; (c) the volume fraction of the magnetically ordered region, derived from the amplitude of the fast relaxing signal; (d) comparison between the present results (red symbol) and those from (Ga,Mn)As12 in a plot of the relaxation rate (which is proportional to the individual ordered moment size multiplied by the moment concentration) versus Tc (which is a measure of the effective average ferromagnetic interaction). A factor 4/3 is multiplied to the parameter a to adjust the difference from the simple exponential decay rate Λ adopted in ref. 12. The good agreement implies that the plotted systems have common mechanisms for their ferromagnetism.
Figure 4: μSR time spectra in zero field. 
ZF μSR time spectra in Li1.1(Zn0.95Mn0.05)As at T=20 K (open circles). The solid line represents the best fit to equation (1). The black and green broken lines show the first and the second terms, respectively, of this fitting function.
Figure 5: μSR time spectra in the WTF. 
μSR time spectra in the WTF of 30 Oe in Li1.1(Zn0.95Mn0.05)As. The oscillation amplitude corresponds to the paramagnetic volume faction.
right
